Wave forecast information is fundamental of the safety operation of working vessels for port construction. There are two distinct methods for wave forecasting. One is a numerical medel of wind and wave interaction. The other is an empirical model based on a statistical relationship between the weather and the wave data obtained in the past. The latter method utilizes simple regression equations or relationships and does not need sophisticated knowledge in the process of practical computation. However, because of the lack of continuous field wave data, detailed study on latter method has been difficult, and only few reports have been presented.
　In this report, the reliability of the three kind of empirical forecasting models (multiple regression model, multichannel autoregressive model and the method with the use of contingency table) is examined through simulation based on 4-year wave data and weather maps. The AIC (Akaike's Information Criterion) is introduced to select vital variables to determine wave heights at specific locations. The examination shows that the reliability of short-term forecasts by these mtthods is satisfactory for practical use of errors are allowed to a certain extent.